686 Inorganic Chemistry, Vol. 11, No. 4, 1972
ANTHONYADINAND
CONTRIBUTION FROM THE INSTITUTE FOR
JAMES
H. ESPENSON
ATOMICRESEARCH AiiD DEPARTMEST O F CHEMISTRY, IOWASTATE UNIVERSITY, AMES,IOWA50010
Kinetics of Some Oxidation Reactions of Diaquocob (1I)aloximela BY ANTHONY ADIN
AND
JAMES H. E S P E N S O P Ib
Received June 28, 1971 Cobalt(II1) complexes of the family Co(NHa)6X2+are rapidly reduced by Co(dmg)a(HeO)?resulting, by transfer of group X, in the formation of Co(dmg)?(HzO)(X). Kinetic measurements were made on these reactions, which follow a second-order rate expression after correction for the decomposition of the Co(I1) reactant. The reactivity of the family of complexes is discussed in terms of the stabilities of the various X-bridged inner-sphere transition states.
Introduction Considerable interest in the bis(dimethyg1yoximato) complexes of cobalt has been expressed in the last few years with the discovery that their chemical properties closely resemble those of vitamin B12. Schrauzerz has recently reviewed the chemistry of these dmg complexes, often called “cobaloximes,” pointing out their similarities to vitamin Blz. We have studied the rates of oxidation of ColI (dmg)z( H z O )by ~ a related series of oxidants. We felt it important to examine a Blz model system that might provide some assessment of the rates and mechanism of the Co(I1)-Co(II1) interconversions. The particular system chosen was the reduction of Co(II1) complexes of the type Co(NH&X2+ by CoII(dmg)2(Hz0)2. The likely oxidation products are Co1I1(drng)z(HzO)X if the reactions proceed by the inner-sphere mechanism or [Co1I1(dmg)2(H,0),]+ from the outer-sphere mechanism Both of these Co(II1) products are kinetically inert toward substitution, 4 , 5 which adds to the attractiveness of this system because the ease of their isolation and characterization permits a positive determination of the mechanism of the redox reactions. Aqueous solutions of aquocob (11)aloxime are stable in the absence of acids and oxygen. In the present work the reactions were carried out with the strict exclusion of oxygen; the complex is reported to be a partially reversible oxygen carrier.6 The decomposition in the presence of hydrogen ion, with concomitant loss of dmg, must be contended with in any aqueous reaction system. The decomposition process can be controlled, a t least for a short time, by appropriate buffers, although i t limits the oxidizing systems that can be studied to nonhydrolyzable compounds such as the pentaamminecobalt(II1) ions. A related reaction is the oxidation of pentacyanocobaltate(I1) ions by the same family of cobalt(1II)ammine c ~ m p l e x e s . ~ Experimental Section Materials.-DiaquocoblI1)aloxime was prepared by t h e method of Schrauzers in an apparatus which allowed the formation, filtration, washing, and drying of t h e product under nitrogen. T h e complex was stored over calcium chloride in a vacuum ( 1 ) (a) Work performed in the Ames Laboratory of the U. S . Atomic Energy Commission; Contribution No. 3048. (bj Fellow of the Alfred P. Sloan Foundation, 1968-1970. (2) G. K . Schrauzer, Accounts Chem. Res., 1, 97 (1968). ( 3 ) The monoanion of dimethylglyoxime is abbreviated dmg. (4) D. N. Hague and J. Halpern, I n o r g . Chew., 6, 2059 (1967). (5) H. G. Tsiang and W. K. Wilmarth, ibid., ‘7, 2835 (1968). (6) G. N. Schrauzer and L. P. Lee, J . Amer. Chrm. Soc., 9 2 , 1551 (1970). (7) J. P. Candlin. J. Halpern, and S.Nakamura, ibid.,86, 2517 (1963). (8) G. N. Schrauzer, Inorg. Sp., 11, 61 (1968).
desiccator. Purity was checked by analysis of t h e Co(I1) released by heating a sample of the solid with concentrated nitric acid. Anal. Calcd for Co(CaN202Hi)?(H?O)z: Co, 18.2. Found: Co, 17.6. This complex shows a distinctive absorption maximum a t X 460 nm in organic or aqueous solvents. Aqueous solutions are stable for a t least 1 hr after rapid initial reactions with traces of oxygen or acid. T h e extinction coefficient a t 460 nm was found t o be 3.84 X lo3 M - 1 cm-1, measured from the decrease in absorbance upon addition of a known but stoichiometrically deficient quantity of Co(NH3)jBrZ+,with allowance made for t h e absorbance of t h e products. The Co(II1) complexes Co(dmg)%Br(HnO) and Co(dnig)ZCl(HzO) were prepared by the method of Ablov and Samus.g The acidopentaamminecobalt (111) complexes were prepared by references cited in earlier publications and were generally converted t o the perchlorate salts except for Co(NH3)jBr2+which was used as the bromide.10,l i Lithium perchlorate was prepared from the carbonate and recrystallized twice. ilnalytical grade sodium acetate, sodium bromide, and glacial acetic acid were used without further purification. Conductivity water was used throughout. Rate Measurements.-The reactions between Co(KHs)sX*+ complexes and Co(dmg)?(H?O)zwere so rapid t h a t the use of t h e stopped-flow method was required. T h e instrument used has been previously described . l o T o avoid interference from the acid decomposition of Co(dmg)z(H20j,, the reactions were carried out in a n acetic acid-acetate buffer. Even at p H 5 , however, t h e acid decomposition slowly occurs. To prevent deterioration of the solution, only the sodium acetate component of the buffer mas added t o the Co(dmg)z(H20)2 stock solution, effectively preventing any decomposition. Only t h e acetic acid buffer component was present in the C O ( S H ~ ) ~ Xsolutions *+ which also served to arrest decomposition of the complexes by base hydrolysis which can be appreciable over long times at p H 5 . T h e kinetics were followed using the 460-nm absorption maximum of Co(dmg)z(HzO)zin most experiments. All runs were carried out a t 24.9 0.1’. T h e ionic strength was maintained a t 0.1 M using lithium perchlorate, and the hydrogen ion concentration was regulated by an acetate buffer with a total concentration of 0.010 M. The p H of each solution from the kinetic runs was measured using a Beckman expanded-scale pH meter. Product Analysis.-Blank experiments on mixtures of CozCand the Co(II1) complexes deemed the most likely products such as Co(dmg)z(H?O)X and [ C 0 ( d m g ) ~ ( H ~ 0were ) ~ ] +carried out. It was found t h a t adequate separations could be achieved using a chromatographic technique with Dowex 50W-X8 cation-exchange resin. The resin was in the H+ form for all complexes except for X- = F- and KCS-, where Na+ resins were necessary. In the blank experiments it was demonstrated t h a t the cationic species e o z + , [ C 0 ( d m g ) ~ ( H ~ 0+, ) ?and ] excess C O ( S H ~ ) ~were X ~ +held in the resin column while the uncharged Co(dmg)t(HzO)X species was washed through the resin by water. T h e wavelength maxima of t h e authentic Co(dmg)z(H?O)X complexes were measured. Their molar absorptivities were also determined from a cobalt analysis made by prior reduction t o eoz’ with excess Cr(I1) and a spectrophotometric analysis with thiocyanate. These Co(dmg)?(HzO)X complexes have Amax 240-250 nm, with molar absorptivities determined in this manner of (2-2.5) X 104 AM-1 cm-1. The solutions from the kinetic
+
~~
(9) A. V. .4blov and N. M. Samus, Russ. J . I ~ o v g Lhem., . 4, 410 (1960). (10) R. T. Wang and J. H . Espenson, J . Ameu. Chem. Soc., 93, 380 (1971). (11) 0. J. Parker and J. H. Espenson, ibid., 91, 1968 (1969).
Inorganic Chemistry, Vol. 11, No. 4 , 1972 687
DIAQUOCOB (II)ALOXIME studies were passed through t h e resin column, and t h e spectra of the water eluents were determined to establish t h e identity and t h e yield of any Co(dmg)z(HzO)X product.
Results Decomposition of Co (dmg)2(H20)2.-Cob (11)aloxime is subject to acid decomposition as shown in the reaction Hi0
+
Co(dmg)z(HnO)?: 2 H +
+CO,,~++ 2Hdmg
(I)
The reaction rate is strongly acid dependent; in 0.001 F perchloric acid, for example, i t proceeds to completion a t a rate too high for the stopped-flow method. Decomposition proved to proceed slowly in acetate buffers of pH 4.8-6.1. Reaction I follows a pseudo-first-order rate expression, and is highly pH dependent with half-times between 2.9 sec a t pH 4.9 and 68 sec a t pH 6.0. The data are summarized in Figure 1 in the form of a plot of kdeo VS. PH. 0.5
or second-order rate expression in the form presented previously." For the slower reactions a correction for the simultaneous decomposition of the C ~ ( d r n g ) ~ ( H ~ O ) ~ reactant was applied to the data, subtracting the value of kdeo a t the pH in question from the experimental pseudo-first-order constant. After such correction, all the kinetic data were in accord with the second-order rate expression -d [Co (dmg )Z(HZO)Z] /dt = k2[Co(NHs)6XZ+] [ C o ( d m g ) z ( H ~ O ) ~(1) ]
The kinetic data for each complex are summarized in Table I. The value of k2 proved to be independent of TABLE I KINETICDATAFOR THE REACTION" BETWEEN Co(NH3)6X2+COMPLEXES AND Co(dmg)z(HsO), ,-----Concn ranges, M--X
104 [Co(NHa)sXa+l
106[H+l
k2, M-1
sec-1
0 24-1.4 (3 2 i 0 3) X lo6 e1 0.20-1 3 (1 40 i. 0.08) X lo4 N3 0 20-1 51 (6 4 i.0 3) X lo3 NCS 5-50 0.20-2 2 (2 5 rt 0 . 1 ) X 10' 0.16-0.74 (7 7 2z 1 5 ) X IOzc OHb 14-70 F 48-202 0.09-0 37 10 5 2z 0 . 9 a At 24.9" in HOAc-OAcbuffer (0.01 M )with fi = 0.10 M (LiC104). T h e concentration cited is [Co(NH3)6HzO]tot. Corrected for the protonation of Co(NH3)5OHz+as described in the text. Br
0 4-1 55 0.3-26 5 5.4-51
0. I
0.01
PH
Figure 1 -The apparent first-order rate constant for the decomposition of CoIr(dmg)z(HzO)zas in reaction I a s a function of pH a t 25' in 0 01 M acetate buffer.
In order to correct for reaction I during the slower redox reactions, the decomposition rate was evaluated as a function of pH under conditions identical with those of the main reaction. Reaction Products.-Using authentic samples of the monosubstituted Co(II1) complexes, Co(dmg),(H20)X, and of the diaquo complex, the ion-exchange procedure outlined above was used to demonstrate that, in every case, the redox reaction resulted in transfer of X
+ Co(dmg)2(H20)2+
Co(NH3)5X2+
+ Co(dmg),(HzO)X + 5NH3
CO,,~+
[ H f ] except for the complex Co(NH3)5H203+ as discussed below. The reactions of C O ( N H ~ ) ~ B and ~~+ Co(NH3)5CI2+were also studied t o learn the effect of certain other variables. The rate constant for the former complex proved to be independent of added free bromide ion, and that of the latter was unchanged when ionic strength was raised from 0.1 to 0.2 M , when the reaction was followed a t 265 rather than 460 nm, and when the acetate buffer was omitted. Consistent values oi kz for the slowly reacting Co(NH3)6F2+complex were obtained, despite a substantial correction, up t o SO%, for the decomposition of the Co(dmg)2(HnO)2 reactant. At pH 25 an appreciable proportion of Co(NH3)bK O 3 + is converted to the hydroxo form in accord with the equilibrium Co(NHa)ijHzO'+ = Co(NH3)50HZt H + K, (111)
+
for which the reported equilibrium constant is 6.0 X lo-' M a t 25" and 1.1 = 0.30 M.I2 Because an accurate value of Ka was not available under these present conditions, the data were treated in a fashion which regards K , as an unknown. Assuming both the aquo and hydroxo complexes react with C0(dmg),(H,0)~, the apparent rate constant kz would depend on [H + ] as given by k z ( & 4- [ H t l } = ~ o H K-k, ka,o[H+]
(11)
The exclusive products resulting from these reactions under the conditions of the kinetic studies were the Co (dmg)z(HzO)X species. Kinetic Studies.-Most kinetic experiments were carried out in the presence of a fairly substantial excess of the Co(II1) complex, except in the case of the Co(NH3)6Br2+reaction where the high rate necessitated the use of second-order conditions. The kinetic data were evaluated using the integrated pseudo-first-order
( 2)
It was not possible by varying Ka to make the left-hand side of eq 2 increase regularly with [H+]. This suggests that the reactivity of the aquo complex is negligible relative to that of the hydroxo species, under this assumption expression 3 is obtained. The data conl/kz = l/koH -k [ H + l / K , k o ~
(3)
formed well with this relation, leading to the values OH = 7.7 X l o 2 M-l sec-' and K , = 8.1 X M. (12) R C Splinter, S J Hariis and R S Tobias, l n o v g C h e m , 7 , 897 (1968)
688 Inorganic Chemistry, VoL 11, No. 4 , 1972
This value of Ka agrees well with the literature value12 under similar conditions. Discussion The decomposition of diaquocob(I1)aloxime in 0.01 M acetate buffer occurs in a first-order process showing a dependence on [ H + ] slightly greater than first-order (ca. 1.1-1.2 order) between pH 4.5 and 6.1. The likely role of H + in this process is the protonation of both oxygens of the dimethylglyoxime ligands thereby destroying the 0---H---0 hydrogen bonding which lends such great stability to these complexes. Independent evidence exists for the importance of such protonation in CoII1(dmg)zc ~ m p l e x e s .l’4~ The rate of reduction of RC12+ was found t o be the same whether or not acetate was present, suggesting that any interaction between -0.01 M acetate and cob(I1)aloxime is slight. The rate constants summarized in Table I show a wide range (-3 X lo4) in reactivity between the different Co(NH3),X2+ complexes. The very slow reaction rate of Co(NH3)5H203+was referred to previously, and we have also found that Co(NH3)e3+reacts too slowly to measure (;.e., decomposition of Co(dmg)2(HzO)2 is the only process seen). Except for Co(NH3)jOH2+,the product determinations referred t o above establish that all the reactions proceed by an inner-sphere mechanism ; we presume the hydroxo complex reacts in the same fashion. The basis of this claim rests on the assumption that anation of Co(dmg)Z(HzO)z+by X- in a medium similar to the present one does not occur rapidly. The kinetics of the anation reactions were not studied, but inappreciable quantities of the complex Co(dmg)2(Hz0)Xwere found in the control experiments. The order of effectiveness of the halide ions as bridging ligands (Br- > C1- >> F-) is reflected in the respective values of the second-order rate constants or more (13) R D Gillard and G Wilkinson, J Chem Soc , 6041 (1963). (14) A Adiii and J H. Espenson, Chem Commun , 753 (1971)
CAROLYN W. MARICONDI AND BODIEE. DOUGLAS precisely by the relative stabilities of the respective transition states according to Haim.15 Both these observations demonstrate the “soft acid” or “class b” character of Co(dmg)2(H20)2 in the transition states and also presumably in the free complex. Generally, a high rate ratio for reduction of an azido complex compared to the N-bonded thiocyanato complex can’be expected,16provided the reductant exhibits hard-acid character. For example, the rate constant ratio is 1.5 X l o 4 for the reductions of (NH3)5CoN32+ and (NH3)jCoNCS2+by Cr2+.” The ratio for the reduction of the same two complexes by the soft acid Co(CN)j3- is 1.6.’ For CoII(dmg)2(H20)2 the ratio is 26, again indicating a tendency toward soft-acid character. Considering the rate constants found for certain other Co(I1)-Co(II1) reactions, the present reactions appear rather rapid. We note t h a t Co(dmg)z(HzO)2is a low-spin complex,” and the electron is transferred from and into an eg orbital. Indeed i t has been observed18 that C ~ ( d r n g ) ~ ( H ~ O ) reacts very rapidly with Fe(CN)e3-, a t a specific rate > 3 X lo7 M-l sec-l. This reaction produces a new anionic species which we presume to be [(NC);FeCNC 0 ( d m g ) ~ H ~ 0 ]by ~ - analogy with Co(CN)z3- reaction~.’~ The observation of this reaction and the importance of the inner-sphere mechanism for the Co(NHg)5X2+ reactions necessarily implicate rapid axial ligand exchange for the complex Co(dmg)2(H20)2. One practical benefit of the present reactions should also be noted. They can provide a convenient synthetic method for Co(dmg)z(HzO)X complexes. This is particularly useful for the fluoro complex, which is difficult to prepare by other methods. (15) A . Haim, Inoug. Chem., 7, 1475 (1968). (16) J. H . Espenson, ibid., 4, 121 (1966). (17) J. P. Candlin, J. Halpern, and D. Z. T r i m m , J . Amer. Chenz. Soc., 86, 1019 (1964). (18) A. Adin, unpublished experiments. (19) A. Haim and W, K. Wilmarth, J . Amer. Chem. Soc., 83, 509 (1061).
CONTRIBUTION FROM THE DEPARTMENT OF CHEMISTRY, UNIVERSITY O F PITTSBURGH, PITTSBURGH, PENNSYLVANIA 15213
Importance of Asymmetric Nitrogens to the Circular Dichroism of Alkyl- Substituted Ethylenediaminediacetic Acid Complexes of Cobalt(1II)l BY CARDLYlU W. MARICONDI
AND
BODIE E. DOUGLAS*
Received July 6 , 1971 Four Co( 111) complexes containing N-alkyl-substituted ethylenediaminediacetic acid have been completely resolved, and their absorption and circular dichroism spectra are reported. Two of the complexes, trans- [Co(DMEDDA)en]I and trans- [Co(DEEDDA)en]I (containing, respectively, the N,N’-dimethyl- and N,N’-diethylethylenediamine-L\7,~V’-diacetate ion), have been reported previously. The two corresponding complexes containing oxalate ion in place of ethylenediamine are new. Both ethylenediamine complexes give very similar absorption and circular dichroism spectra, as do the two oxalate complexes T h e C D peak intensities of all the complexes, however, are less than half those of the corresponding unsubstituted-EDDA complexes. This reduction in intensities is attributed to a lower contribution to the asymmetry of the complexes from nitrogens in the N-alkyl-substituted EDDA’s as compared to unsubstituted E D D h
Introduction I n the course of recent studies to determine factors responsible for variations in intensities of circular dichroism (CD) bands within and between series Of re(1) This work was supported b y a research grant (GILI-10829) from the Division of General Medical Studies, U. S. Public Health Service.
lated complexes, a series of cobalt(II1) complexes containing the tetradentate ligand ethylenediamine-hT,N’diacetate (EDDA) was p r e ~ a r e d . ~The . ~ geometrical (2) J. I. Legg, D . U‘. Cooke, and B. E. Douglas, Inorg. Chem., 6 , 700 (1967). (3) C. W. Van S a m and B. E. Douglas, ibid., 8 , 116 (1969).
